符号(正式)
人工神经网络
回归分析
回归
机器学习
人工智能
工程类
还原(数学)
预测建模
计算机科学
统计
数学
几何学
程序设计语言
作者
Shaowu Feng,Xingyue Sun,Gang Chen,Hao Wu,Xu Chen
标识
DOI:10.1016/j.ijfatigue.2023.107962
摘要
This study employs machine learning models assisted by symbol regression to achieve satisfactory corrosion fatigue life prediction for T91 steel and 316L stainless steel (SS) used in 4th generation advanced nuclear power plants. Symbol regression features improve the performance of all considered machine learning models. The artificial neural network (ANN) model shows the most significant improvement of 22 % decrease in RMSE regarding to the model without symbol regression features. Additionally, the well-trained ANN model is transferred to predict 316L SS with a 50 % reduction in training samples, highlighting its potential for more efficient model training and deployment.
科研通智能强力驱动
Strongly Powered by AbleSci AI